ServiceNow's twice-yearly family releases are a gift and a treadmill. A gift, because the platform improves relentlessly. A treadmill, because falling behind N-1 means losing support coverage — and catching up two releases at once is genuinely painful. The teams that stay current without drama run upgrades as a factory, not a project.
Habit one: automated regression, or nothing
The cost of an upgrade is dominated by testing. Manual regression means weeks of coordinated user time twice a year — which is why it quietly stops happening, which is why upgrades become scary. Automated Test Framework (ATF) coverage of your critical paths converts testing from a campaign into a button. We target coverage of every flow that would page someone if it broke; typically 150–400 tests for an enterprise instance.
Habit two: know your debt before the release notes drop
Every customization is a potential collision with the new release. Instances that track their deviations — what was changed, why, and who owns it — can read release notes against a known inventory. Instances that don't, rediscover their customizations by breakage.
Habit three: rehearse in sub-production, always
Clone down, upgrade the clone, run the ATF suite, fix what surfaced, then schedule production. The rehearsal usually takes days; skipping it converts those days into an outage lottery.
Habit four: upgrade in quiet hours you've already agreed
Standing upgrade windows negotiated once a year beat window negotiations twice a year. Our AMS clients pre-agree calendars; most production upgrades complete inside a five-day cycle from clone to done, with the business noticing nothing but the new features.
The compounding payoff
Staying current isn't just support hygiene. Now Assist capabilities, performance improvements and security patches all land in current releases first. The instances that upgrade routinely compound; the ones that defer pay interest. The factory model is how routine becomes real.